RAISE 框架：如何规范报告教育领域的 AI 研究

Artificial intelligence (AI) is becoming increasingly visible in education. Teachers are using generative AI for lesson planning and feedback, while educational institutions are trialling adaptive learning platforms, exploring AI-supported tutoring systems (Létourneau et al., 2025) and adapting assessments. Alongside this growth, educators are also encountering an increasing number of claims about what AI can supposedly achieve for learning, attainment and efficiency.

This blog post introduces the Reporting AI Studies in Education (RAISE) framework (Allison, 2026), which aims to improve how AI research is reported in educational settings so that evidence is transparent, interpretable and relevant to practice.

The problem is that these claims can be difficult to evaluate. Useful AI research needs to provide more than technical descriptions or headline outcomes. Educational technology research does not always provide enough information about how a tool was implemented, what role teachers played, or whether positive outcomes can realistically be transferred into different educational settings. As a result, schools and universities can find themselves making important decisions based on incomplete evidence. Without clear reporting, educators are left with important unanswered questions. Who were the learners involved? What educational problem was the AI intended to address? How much teacher support was required? Was the technology used for feedback, revision, tutoring or assessment? Were there safeguards to address inaccurate outputs, bias or accessibility concerns?

‘Schools and universities can find themselves making important decisions based on incomplete evidence.’

Why reporting standards matter in educational AI

Reporting standards in AI research help readers judge the quality, credibility and applicability of findings. In education, this is particularly important because teaching and learning are highly contextual. An AI system that appears effective in one setting may not produce similar outcomes elsewhere.

In systematic literature reviews, PRISMA has often been used for accurate reporting. In medicine, frameworks such as CONSORT have long been used to improve transparency in clinical trials. In education, however, reporting practices around AI remain inconsistent despite the rapid expansion of AI-related research. RAISE was developed in response to this gap, recognising that educational AI research requires reporting that captures both technical and pedagogical dimensions of implementation.

What the RAISE framework does

The Reporting AI Studies in Education (RAISE) framework contains 30 checklist items within 10 domain areas. Rather than focusing only on technical performance, the framework encourages researchers to explain how AI tools are situated within teaching and learning processes. This includes reporting information about educational context, pedagogical rationale, ethical considerations, implementation decisions and the role of educators within AI-supported activities.

For example, educational context reporting requires researchers to explain who the learners were, where the intervention took place and what educational challenge was being addressed. Pedagogical rationale requires explanation of why AI was selected and how it aligns with learning objectives. Reporting educator involvement encourages researchers to describe how teachers supported, monitored or mediated learner interactions with AI systems. This is important because AI tools do not operate independently from educational practice; effectiveness depends on context and human judgment.

RAISE is also notable because it encourages more explicit discussion of risks and ethics through a fillable template matrix (figure 1), while also providing good and bad reporting examples. Current debates about AI in education often focus on efficiency and innovation, but practitioners are increasingly concerned about issues such as privacy, bias, accessibility and learner dependency. These are not abstract questions. Teachers are already navigating concerns about responsible assessment design and student overreliance on generative AI tools. Transparent reporting helps practitioners assess these risks more critically.

Figure 1: RAISE ethics and risk matrix (Allison, 2026)

For practitioners, RAISE can help shift attention away from simple claims that an AI tool ‘works’ towards more meaningful questions about under what conditions it works, for whom, and with what trade-offs, therefore aiding practitioners in trying to make evidence-informed decisions.

AI is likely to remain an important part of educational innovation. However, innovation requires evidence that is transparent, interpretable and relevant to practice. RAISE is a step forward to help make AI evidence more meaningful for those responsible for educational decision-making.

References

Allison, J. (2026). RAISE the standard: A framework for transparent reporting of artificial intelligence studies in education. Journal of Educational Computing Research, 64(1), 3–15. https://doi.org/10.1177/07356331251377430

Létourneau, A., Deslandes Martineau, M., Charland, P., Karran, J. A., Boasen, J., & Léger, P. M. (2025). A systematic review of AI-driven intelligent tutoring systems (ITS) in K-12 education. npj Science of Learning, 10(1), 29. https://doi.org/10.1038/s41539-025-00320-7